# =============================================================================
# HSCI 841 Qualitative Research Methods and Analysis in Public Health  -  Lesson 4:
#   Qualitative Data Collection
# Answer key for the in-lesson R activities
# Data file(s): the 20 transcripts in term projects/HSCI_841/transcripts/
#   (P01_Maya.txt ... P20_Frank.txt; all of them are in HSCI_841_loneliness_data.zip)
# Packages: readtext, quanteda, dplyr  (install once with
#   install.packages(c("readtext", "quanteda", "dplyr")))
# Reproduces every code block in the lesson.
#
# Paths are relative to the course repository root. Set your working directory there
# (Session > Set Working Directory > Choose Directory, or open the course .Rproj) before running.
# =============================================================================

# NOTE (answer key): the transcripts contain typographic dashes and curly quotation marks. If the R
# session is not running in a UTF-8 locale (plain Rscript on a server, for example), those characters
# are escaped rather than read, which pollutes word-frequency output. RStudio on Windows, macOS and
# Linux is already UTF-8; this line only matters for a bare Rscript run.
if (!isTRUE(l10n_info()[["UTF-8"]])) invisible(suppressWarnings(Sys.setlocale("LC_CTYPE", "C.UTF-8")))

# ==== Section 3: Reading the Corpus into R / Read the transcript corpus and summarise word counts ====

# Read all 20 loneliness transcripts as a corpus
library(readtext)
library(quanteda)
library(dplyr)

# NOTE (answer key): path corrected from "../term projects/..." to the repository-root convention.
transcript_dir <- "term projects/HSCI_841/transcripts"

# readtext::readtext() reads every .txt file in a folder into a tibble
# with one row per file, a doc_id column, and a text column
loneliness_texts <- readtext(file.path(transcript_dir, "*.txt"),
                            docvarsfrom = "filenames",
                            dvsep = "_",
                            docvarnames = c("pid", "pseudonym"))

# Turn it into a quanteda corpus object
loneliness_corpus <- corpus(loneliness_texts)

# Words per transcript
word_counts <- ntoken(loneliness_corpus, remove_punct = TRUE)

# Build a tidy summary
summary_df <- tibble(
  pid = docvars(loneliness_corpus, "pid"),
  pseudonym = docvars(loneliness_corpus, "pseudonym"),
  words = as.integer(word_counts)
) %>% arrange(desc(words))

print(summary_df, n = 20)
summary(summary_df$words)
# Median, IQR, min, max give a sense of the spread in transcript length

# ---- Reading the summary ----
cat("Documents:", ndoc(loneliness_corpus), " total words:", sum(summary_df$words), "\n")
cat("Longest:", summary_df$pseudonym[1], summary_df$words[1], "words;",
    "shortest:", summary_df$pseudonym[nrow(summary_df)],
    summary_df$words[nrow(summary_df)], "words\n")
cat("Median", median(summary_df$words), "words, IQR",
    paste(quantile(summary_df$words, c(0.25, 0.75)), collapse = " to "), "\n")
# The spread is a data-collection finding, not a nuisance: the shortest interviews are with the
# youngest participant (P12 Tyler) and with P10 Daniel, who was guarded about a topic he has not
# discussed with his family. Interview length is itself evidence about rapport and elicitation.

# Interviewer and participant talk, separately: a simple check on who is doing the talking.
turn_counts <- lapply(list.files(transcript_dir, pattern = "\\.txt$", full.names = TRUE), function(f) {
  lines <- readLines(f, warn = FALSE)
  data.frame(file = basename(f),
             interviewer_turns = sum(grepl("^I:", lines)),
             participant_turns = sum(grepl("^P( \\(continued\\))?:", lines)),
             participant_words = sum(lengths(strsplit(grep("^P( \\(continued\\))?:", lines, value = TRUE), "\\s+"))))
})
turn_counts <- bind_rows(turn_counts)
print(turn_counts)
cat("Participant words as a share of all words:",
    round(100 * sum(turn_counts$participant_words) / sum(summary_df$words)), "%\n")
# A guide that works keeps the interviewer's share small. Here the participants do most of the
# talking, which is what a semi-structured guide is for.